# shinkuan/Akagi

支持雀魂、天鳳、麻雀一番街、天月麻將，能夠使用自定義的AI模型實時分析對局並給出建議，內建Mortal AI作為示例。 Supports Majsoul, Tenhou, Riichi City, Amatsuki, with the ability to use custom AI models to analyze games in real time and provide suggestions. Comes with Mortal AI as a built-in example.

Repository: https://github.com/shinkuan/Akagi
Canonical: https://ross.abutalabs.com/products/akagi
Homepage: https://akagi.shinkuan.me/
Language: Rust
License: Apache-2.0
License Family: permissive
Topics: game-ai, japanese-mahjong, mahjong, mahjong-ai, mahjong-bot, majsoul, majsoul-ai, majsoul-plus, riichi-mahjong
Last push: 2026-08-25T03:51:54+00:00

## Health v2 (maintenance only)
Score: 93/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 99, longevity 69
- inputs: {"age_days": 971, "days_push": 8, "days_rel": 8, "gap_med": 2, "n_releases_24m": 46}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1007, forks 123 (observed 2026-08-28T04:03:12.344155+00:00)

## What it is
Akagi is a real-time mahjong AI assistant desktop app (Rust + Tauri) that observes game connections from Mahjong Soul, Tenhou, Riichi City, and Amatsuki, rebuilds the table state, and shows shanten, waits, deal-in risk, and AI discard recommendations in a HUD. It ships with the Mortal AI built in and supports custom mjai-compatible bots and optional cloud inference.

## Use cases
- get real-time discard suggestions while playing riichi mahjong
- analyze my Mahjong Soul games live with an AI
- see shanten, waits, and deal-in risk during a match
- improve at riichi mahjong by learning from AI recommendations
- run a custom mjai bot against live mahjong games
- review three- and four-player mahjong games with AI analysis

## When to choose
- you play on Mahjong Soul, Tenhou, Riichi City, or Amatsuki and want live in-game analysis
- you want a single desktop app with a built-in AI (Mortal) and no complex setup
- you want to plug in your own mjai-compatible bot or stronger cloud inference
- you want a passive observer that reads game data without modifying the game client

## When to avoid
- you plan to use it in ranked or competitive play - it may violate the games' terms of service and risk account bans
- you play mahjong clients not on the supported list
- you need offline batch analysis of game logs rather than live assistance
- you need a fully open AI model - the strongest inference options may be cloud-based

## Facets
- artifact type: application
- maturity: active
- function: game, machine-learning, gui, websocket
- domain: gaming-tools, machine-learning
- platform: windows, rust
- tags: mahjong, mahjong-ai, riichi-mahjong, majsoul, tenhou, mjai, game-assistant, real-time-analysis, tauri, hud-overlay, real-time, desktop, macos, linux

## Member repositories
- shinkuan/Akagi (main) score 93

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:12.344155+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:12:41.905770+00:00, confidence not recorded.
  - readme: https://github.com/shinkuan/Akagi (fetched 2026-08-28T04:03:12.344155+00:00, sha 38294342f41f)
  - homepage: https://akagi.shinkuan.me/ (fetched 2026-08-29T13:12:36.168997+00:00, sha 5f39b0501c96)
  - site_page: https://akagi.shinkuan.me/docs (fetched 2026-08-29T13:12:36.178851+00:00, sha 8031d91b2bd6)
- Data as of 2026-08-30T08:39:29.467469+00:00.